Research map / Optimization / Nonconvex and smooth optimization
Variants of RMSProp and Adagrad with Logarithmic Regret Bounds
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use SC-Adagrad with a non-increasing damping factor for strongly convex losses | Adagrad · Strongly convex online convex optimization | regret growth bound: O(\sqrt{T}) -> O(\log(T)) | Helped |
| Use SC-RMSProp with a non-increasing damping factor and decaying step size | RMSProp · Strongly convex online convex optimization | regret growth bound: logarithmic regret bound | Helped |
| Choose β_t within the analyzed time-varying range | RMSProp · Convex online convex optimization | regret growth bound: O(\sqrt{T}) | Helped |
| Use RMSProp (Ours) instead of original RMSProp with constant β_t = 0.9 | RMSProp · Strongly convex online softmax regression | regret | Helped |
| Use SC-Adagrad or SC-RMSProp instead of the other compared methods | Strongly convex online softmax regression · Softmax regression | regret | Helped |
| Use RMSProp (Ours) with time-varying β_t instead of the original constant β_t = 0.9 | RMSProp · CIFAR10 CNN | training objective | Helped |
| Use RMSProp (Ours) or SC-Adagrad | Compared optimization methods · CIFAR10 CNN | test accuracy | Helped |
| Use SC-RMSProp | Compared optimization methods · CIFAR100 and MNIST CNN | test accuracy | Mixed |
| Use RMSProp with constant β_t = 0.9 instead of the analyzed time-varying weighting scheme (ablation) | RMSProp · CIFAR10 CNN | convergence | Hurt |
| Replace SC-Adagrad's damping with the small numerical-stability value used by standard adaptive methods (ablation) | SC-Adagrad · CNN training | training objective | Hurt |
Closest papers by meaning
- Small-loss Adaptive Regret for Online Convex Optimization (ICML 2024 · similarity 0.91)
- Fast Rates in Stochastic Online Convex Optimization by Exploiting the Curvature of Feasible Sets (NeurIPS 2024 · similarity 0.89)
- Improved Regret for Bandit Convex Optimization with Delayed Feedback (NeurIPS 2024 · similarity 0.88)
- Online Inverse Linear Optimization: Efficient Logarithmic-Regret Algorithm, Robustness to Suboptimality, and Lower Bound (NeurIPS 2025 · similarity 0.88)
- From Non-Convex to Strongly Convex: Curvature-Adaptive FTPL for Online Optimization (NeurIPS 2026 · similarity 0.88)
- A Reduction from Delayed to Immediate Feedback for Online Convex Optimization with Improved Guarantees (NeurIPS 2026 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.